Glossary · Anthropic

Model Context Protocol (MCP)

MCP is an open standard for connecting AI models to tools, data, and APIs. Definition, ecosystem, and why it matters for AI interoperability.

By Kadin Nestler · May 28, 2026 · Updated May 28, 2026
On this page

What MCP standardizes

Before MCP, every LLM vendor had its own way of describing tools, resources, and prompts to the model. A tool built for Claude tool use did not work with OpenAI function calling without a rewrite. MCP defines a vendor-neutral protocol: an MCP server exposes resources (read-only context), tools (callable functions), and prompts (templates). Any MCP client — Claude Desktop, Cursor, Cline, Continue, ChatGPT desktop — can connect to that server and use it.

The MCP ecosystem in 2026

  • Official servers from Anthropic for filesystem, GitHub, Slack, Google Drive, Postgres, Puppeteer, and more.
  • Hundreds of community servers covering every common SaaS app, database, and developer tool.
  • Major vendors (OpenAI, Microsoft, Google) have announced or shipped MCP client support.
  • IDEs (Cursor, VS Code Copilot, Cline) treat MCP servers as first-class plugins.

Why MCP matters

For developers: write a tool once, use it everywhere. For businesses: pick AI vendors based on quality and cost, not on which one happens to have the integrations you need. For the ecosystem: the equivalent of HTTP for AI tools — a common substrate that lets the rest of the stack innovate. MCP is one of the few standards in the AI space that has achieved meaningful cross-vendor adoption in under 18 months.

Practical implications for SMBs

If your AI vendor is on top of MCP, switching them out is cheaper. If they are not, you are locked into their proprietary integration layer. Ask: "Do you publish your tools as MCP servers, and can I run them locally if I want to host them myself?" The answer reveals how portable the work they ship you actually is.

What it means for your business

MCP is the closest thing the AI tool ecosystem has to a universal connector. Vendors aligned with the standard are bets on portability; vendors building proprietary alternatives are bets on lock-in.

  • Claude Agent SDK — The Claude Agent SDK is Anthropic's open-source framework for building production agents on top of Claude. Definition, capabilities, and when to use it.
  • Tool Use — Tool use is when an LLM calls external APIs, databases, or code on its own. Definition, function calling, and how it powers AI agents.
  • Claude Code — Claude Code is Anthropic's terminal-based AI coding agent — reads your repo, runs commands, edits files, and ships PRs. Definition, pricing, and use cases.
  • AI Orchestration — AI orchestration is the layer that coordinates LLM calls, tools, and data into a working application. Definition, top frameworks, and how to choose.
  • AI Agent — An AI agent is an LLM-driven program that uses tools to complete tasks autonomously. Definition, architecture, and real SMB examples.
← Back to the full glossary